From Project Mahasen

Mahasen Project Introduction

In modern day computing the volume of the data that applications have to handle is increasing rapidly.For an instance The UK e-Science programme has funded the DAME project a consortium analyzing sensor data generated by Rolls Royce aero-engines. It is estimated that there are around 100,000 Rolls Royce engines currently in service. Each trans-Atlantic flight made by each engine generates about a Gigabyte of data per engine – from pressure, temperature and vibration sensors.So grid middle-ware can be used to find solutions which are scalable , efficient and reliable for such large volumes of data management.

Storage Broker is a Data Grid Management System (DGMS) which manages logically distributed file system. The architecture of an SRB is designed in such a way where it presents the user with a single global name-space while handling a distributed file system.
It is a network of storage servers that plays the dual purpose of a meta-data catalog and a storage server. The network of servers are fault tolerant, scalable and acts as a one logical single storage server.

Storage & nbsp;Broker provides the uniform API to retrieve and storage data in a physically distributed storage system.Meta data catalog can assist in making the object accessible while querying the objects stored using the meta data defined at the storing phase.Meta data can be either user defined or system defined which is calibrated with the application.

When we are dealing with huge amount of data, it is required to have a meta-data management system in order to handle data efficiently. As we are going to develop a Distributed Storage Broker, we also do concern on how to manage those data efficiently. Some of the major facts that we are going to research on are
to come up with a good mechanism to efficiently organize and maintain large directories which are having large number of files.
to develop a method for high meta-data performance for large scale file systems
a mechanism to provide high meta-data performance for mixed workload generated by users Which are common to any meta-data managed storage system.